path: root/gs/lib/pf2afm.ps

%!
% This is a PostScript program for making an AFM file from
% PFB / PFA and (optionally) PFM files.
%
% Written in BOP s.c., Gda\'nsk, Poland
% e-mail contact: B.Jackowski@GUST.ORG.PL
% version 0.5 (18 XII 1997)
% version 0.55 (11 III 1998) -- unlimited number of chars in a font
% version 1.00 (27 III 1998) -- scanning PFM subdirectory added,
%                               code improved; version sent to LPD
% version 1.01 (1 II 2000)   -- message changed

% Usage:
%   gs [-dNODISPLAY] -- pf2afm.ps disk_font_name
%
% The result is written to the file disk_font_name.afm, provided such
% a file does not exist; otherwise program quits.
%
% The font can be either *.pfa or *.pfb; if no extension is supplied,
% first disk_font_name.pfb is examined, then disk_font_name.pfa.
% Moreover, if there is a *.pfm file in the same directory or in the
% subdirectory PFM, i.e., disk_font_name.pfm or PFM/disk_font_name.pfm,
% kern pairs from it are extracted, as well as additional font
% parameters, usually absent from Type 1 fonts.

% Tribute:
% $Id$
% The program is based on James Clark's <jjc@jclark.uucp> printafm.ps
% (with alterations by d.love@dl.ac.uk and L. Peter Deutsch) from
% Ghostscript 5.10 distribution.

/onechar 1 string def
/edef {exch def} def
/WinAnsiEncoding dup /Encoding findresource def

% charnumber print-charname -
% prints the name of the encoded character
/print-charname {
  PFMCharSet 0 eq {
    WinAnsiEncoding
  } {
    PFBencoding
  } ifelse
  exch get =string cvs dup
  (.notdef) eq {
    /was.notdef true def
  } if
  print.to.ofi ( ) print.to.ofi
} def

/printquit {print flush quit} def

% redirecting GS output to ``ofi'' file
/eolch (\r\n) def
/=only.to.ofi {ofi exch write=only} def          % replaces GS's `=only'
/print.to.ofi  {ofi exch writestring} def        % replaces `print'
/=to.ofi { =only.to.ofi eolch print.to.ofi } def % replaces `='

% read and skip: byte, short, word, double and long
/readb-p {currPFMfile read not {(Unexpected EOF\n) printquit} if} def
/readw-p {readb-p readb-p 256 mul add} def
/reads-p {readw-p dup 32768 ge {65536 sub} if} def
/readd-p {readb-p readb-p readb-p readb-p 256 mul add 256 mul add 256 mul add} def
/readl-p /readd-p load def % double word is, in fact, long integer in GS
/skipb-p {readb-p pop} def
/skipw-p {skipb-p skipb-p} def
/skips-p /skipw-p load def
/skipd-p {skipb-p skipb-p skipb-p skipb-p} def
/skipl-p /skipd-p load def
/skipa-p { {skipb-p} repeat} def

% PFMfile readPFMheader -
% defines currPFMfile, PFMExtMetricOffset, PFMPairKernTableOffset

/readPFMheader {
  currPFMfile bytesavailable
  % ---------------
  % PFM MAIN HEADER
  % ---------------
  skipw-p % PFM: version
  readd-p % PFM: size (size is dword, not word as the documentation says)
  ne {(Wrong file size\n) printquit} if
  60 skipa-p  % PFM: copyright
  skipw-p % PFM: Type
  skipw-p % PFM: Points
  skipw-p % PFM: VertRes
  skipw-p % PFM: HorizRes
  skipw-p % PFM: Ascent
  skipw-p % PFM: InternalLeading
  skipw-p % PFM: ExternalLeading
  skipb-p % PFM: Italic
  skipb-p % PFM: Underline
  skipb-p % PFM: Stikeout
  skipw-p % PFM: Weight
  readb-p % PFM: CharSet
    /PFMCharSet edef
  skipw-p % PFM: PixWidth
  skipw-p % PFM: PixHeight
  skipb-p % PFM: PitchAndFamily
  skipw-p % PFM: AvgWidth
  skipw-p % PFM: MaxWidth
  skipb-p % PFM: FirstChar
  skipb-p % PFM: LastChar
  skipb-p % PFM: DefaultChar
  skipb-p % PFM: BreakChar
  skipw-p % PFM: WidthBytes
  skipd-p % PFM: Device
  skipd-p % PFM: Face
  skipd-p % PFM: BitsPointer
  skipd-p % PFM: BitsOffset
  % here we assume that it is a PostScript font, i.e., it always uses
  % the extended width table, therefore the normal width table is empty
  % -------------
  % PFM EXTENSION
  % -------------
  skipw-p % PFMEX: SizeFields
  readd-p % PFMEX: ExtMetricOffset
    /PFMExtMetricOffset edef
  skipd-p % PFMEX: ExtentTable
  skipd-p % PFMEX: OriginTable
  readd-p % PFMEX: PairKernTable
    /PFMPairKernTableOffset edef
  skipd-p % PFMEX: TrackKernTable
  skipd-p % PFMEX: DriverInfo
  skipd-p % PFMEX: Reserved
} def

% requires that currPFMfile, PFMExtMetricOffset are defined
% readPFMExtMetric -
% defines PFMNumberofKernPairs

/readPFMExtMetric {
  currPFMfile PFMExtMetricOffset setfileposition
  skips-p % EXTT: Size
  skips-p % EXTT: PointSize
  skips-p % EXTT: Orientation
  skips-p % EXTT: MasterHeight
  skips-p % EXTT: MinScale
  skips-p % EXTT: MaxScale
  skips-p % EXTT: MasterUnit
  reads-p % EXTT: CapHeight
   /PFMCapHeight edef
  reads-p % EXTT: XHeight
   /PFMXHeight edef
  reads-p % EXTT: LowerCaseAscent
   /PFMLowerCaseAscent edef
  reads-p % EXTT: LowerCaseDescent
   neg /PFMLowerCaseDescent edef
  skips-p % EXTT: Slant
  skips-p % EXTT: SuperScript
  skips-p % EXTT: SubScript
  skips-p % EXTT: SuperScriptSize
  skips-p % EXTT: SubScriptSize
  skips-p % EXTT: UnderlineOffset
  skips-p % EXTT: UnderlineWidth
  skips-p % EXTT: DoubleUpperUnderlineOffset
  skips-p % EXTT: DoubleLowerUnderlineOffset
  skips-p % EXTT: DoubleUpperUnderlineWidth
  skips-p % EXTT: DoubleLowerUnderlineWidth
  skips-p % EXTT: StrikeOutOffset
  skips-p % EXTT: StrikeOutWidth
  readw-p % EXTT: KernPairs
    /PFMNumberofKernPairs edef
  skipw-p % EXTT: KernTracks
} def

% requires that currPFMfile, PFMPairKernTableOffset, PFMNumberofKernPairs are defined
% readPFMExtMetric -
% prints kern pairs table in the AFM format

/readPFMKernPairs {
  currPFMfile () ne {
    PFMdict begin
    PFMPairKernTableOffset 0 ne {
      currPFMfile PFMPairKernTableOffset setfileposition
      readw-p % undocumented kern count (although all remaining structures are
              % explicitly preceded by their sizes); if it were a stable
             % feature, EXTTEXTMETRICS could be skipped
      PFMNumberofKernPairs
%     2 copy = =
      ne {
        (Inconsistent number of kern pairs\n) printquit
      } if
      (StartKernData) =to.ofi
      (StartKernPairs ) print.to.ofi
      PFMNumberofKernPairs =to.ofi
      % ---------
      % MAIN LOOP
      % ---------
      /was.notdef false def
      PFMNumberofKernPairs {
        (KPX ) print.to.ofi
        readb-p  % first  char
        print-charname
        readb-p  % second char
        print-charname
        reads-p  % kern amount
        =to.ofi
      } repeat
      was.notdef {
        (.notdef character ocurred among kern pairs) =
        (you'd better check the resulting AFM file.) =
      } if
      (EndKernPairs) =to.ofi
      (EndKernData)  =to.ofi
    } if
    end
  } if
} def

% alias (for ``compatibility'' with J. Clark):
/printkernpairs /readPFMKernPairs load def

% printcharmetrics -

/printcharmetrics {
  (StartCharMetrics ) print.to.ofi
  /PFBencoding currfont /Encoding get dup length array copy def
  /PFBcharstrings currfont /CharStrings get def
  PFBcharstrings length
  PFBcharstrings /.notdef known { 1 sub } if =to.ofi
  currfont 1000 scalefont setfont
  % checking Encoding array and CharStrings dictionary for
  % the consistency of names
  /was.inconsitent false def
  0 1 255 {
    dup PFBencoding exch get
    PFBcharstrings exch known {
      pop
    }{
%     dup PFBencoding exch get =
      PFBencoding exch /.notdef put % fix Encoding array
      /was.inconsitent true def
    } ifelse
  } for
  was.inconsitent {
    (Encoding array contains name(s) absent from CharStrings dictionary) =
  } if
  % print metric data for each character in PFB encoding vector
  0 1 255 {
    dup PFBencoding exch get
    dup /.notdef ne {
      exch dup printmetric
    }{
      pop pop
    } ifelse
  } for
  % xPFBencoding contains an entry for each name in the original
  % encoding vector
  /xPFBencoding PFBcharstrings length dict def
  PFBencoding {
    xPFBencoding exch true put
  } forall

  /fontiter 0 def
  /TMPFontTemplate (TMP_FONT#000) def
  {
    % NewPFBencoding is the new encoding vector
    /NewPFBencoding 256 array def
    0 1 255 {
      NewPFBencoding exch /.notdef put
    } for
    % fill up NewPFBencoding with names from CharStrings dictionary that
    % are not encoded so far
    /i 0 def
    PFBcharstrings {
      pop
      i 255 le {
        dup xPFBencoding exch known not {
          dup xPFBencoding exch true put
          NewPFBencoding i 3 -1 roll put
          /i i 1 add def
        }{
          pop
        } ifelse
      }{
        pop exit
      } ifelse
    } forall
    i 0 eq {exit} if
    % define a new font with NewPFBencoding as its encoding vector
    currfont maxlength dict /NewTMPfont edef
    currfont {
      exch dup dup /FID ne exch /Encoding ne and {
        exch NewTMPfont 3 1 roll put
      }{
        pop pop
      } ifelse
    } forall
    % compute a unique name for a font to be registered
    /fontiter fontiter 1 add def
    TMPFontTemplate fontiter (000) cvs
    dup length TMPFontTemplate length exch sub exch putinterval
    /TMPFontName TMPFontTemplate cvn def
    NewTMPfont /FontName TMPFontName put
    NewTMPfont /Encoding NewPFBencoding put
    % make this new font the current font
    TMPFontName NewTMPfont definefont 1000 scalefont setfont
    % print metric data for each character in the newly created encoding vector
    0 1 255 {
      dup NewPFBencoding exch get
      dup /.notdef ne {
        exch -1 printmetric
      }{
        pop pop exit
      } ifelse
    } for
  i 255 lt {exit} if
  } loop
  (EndCharMetrics) =to.ofi
} def

% name actual_code normal_code printmetric -

/printmetric {
  (C ) print.to.ofi =only.to.ofi
  ( ; WX ) print.to.ofi
  onechar 0 3 -1 roll put
  onechar stringwidth pop round cvi =only.to.ofi
  ( ; N ) print.to.ofi =only.to.ofi
  ( ; B ) print.to.ofi
  newpath 0 0 moveto
  onechar false charpath flattenpath pathbbox
  newpath
  round cvi /ury edef round cvi /urx edef
  round cvi /lly edef round cvi /llx edef
  ury lly eq {/ury 0 def /lly 0 def} if % normalize degenrated BB
  urx llx eq {/urx 0 def /llx 0 def} if %
  llx =only.to.ofi ( ) print.to.ofi lly =only.to.ofi ( ) print.to.ofi
  urx =only.to.ofi ( ) print.to.ofi ury =only.to.ofi ( ) print.to.ofi
  (;) =to.ofi
} def

/printinfoitem {
  3 1 roll 2 copy known {
    get dup type /stringtype ne { =string cvs } if exch
    print.to.ofi ( ) print.to.ofi =to.ofi
  }{
    pop pop pop
  } ifelse
} def

/printfontinfo {
  (Comment AFM Generated by Ghostscript/pf2afm) =to.ofi
  currfont /FontName  (FontName) printinfoitem
  %
  currfont /FontInfo get
  dup /FullName           (FullName)           printinfoitem
  dup /FamilyName         (FamilyName)         printinfoitem
  dup /Weight             (Weight)             printinfoitem
  dup /Notice             (Notice)             printinfoitem
  dup /ItalicAngle        (ItalicAngle)        printinfoitem
  dup /isFixedPitch       (IsFixedPitch)       printinfoitem
  dup /UnderlinePosition  (UnderlinePosition)  printinfoitem
  dup /UnderlineThickness (UnderlineThickness) printinfoitem
      /version            (Version)            printinfoitem
  %
  (EncodingScheme FontSpecific) =to.ofi
  %
  (FontBBox) print.to.ofi
  currfont /FontBBox get {
    ( ) print.to.ofi round cvi =only.to.ofi
  } forall
  eolch print.to.ofi
  %
  currPFMfile () ne {
    PFMdict
    dup /PFMCapHeight        (CapHeight) printinfoitem
    dup /PFMXHeight          (XHeight)   printinfoitem
    dup /PFMLowerCaseDescent (Descender) printinfoitem
        /PFMLowerCaseAscent  (Ascender)  printinfoitem
  } if
} def

/readPFBfile {
  % make a shot of the actual font directory:
  /oldFontDirectory FontDirectory dup length dict copy def
  isPFB {% defined in `makeafm'
    (r) file true /PFBDecode filter cvx % true is better (see gs_type1.ps)
    mark exch exec
  }{
    (r) file mark exch run
  } ifelse
  cleartomark
  % make a shot of the updated font directory:
  /newFontDirectory FontDirectory dup length dict copy def
  % spot the added font:
  oldFontDirectory  {pop newFontDirectory exch undef} forall
  newFontDirectory length 1 ne {
    newFontDirectory length =
    (Weird PFB file?\n) printquit
  } if
  newFontDirectory {pop} forall
  findfont /currfont edef
} def

/readPFMfile {
  dup () ne {
    (r) file /currPFMfile edef
    10 dict dup /PFMdict edef begin
    readPFMheader
    readPFMExtMetric
    end
  }{
    pop /currPFMfile () def
  } ifelse
} def

% pfmfilename pf[ba]filename filetype printafm -
% where filetype=(a) or (b)

/printafm {
  readPFBfile
  readPFMfile
  (StartFontMetrics 2.0) =to.ofi
  printfontinfo
  printcharmetrics
  printkernpairs
  (EndFontMetrics) =to.ofi
} def

/pfa_pfb_dict <<
  /.pfb /pfbn
  /.pfB /pfbn
  /.pFb /pfbn
  /.pFB /pfbn
  /.Pfb /pfbn
  /.PfB /pfbn
  /.PFb /pfbn
  /.PFB /pfbn

  /.pfa /pfan
  /.pfA /pfan
  /.pFa /pfan
  /.pFA /pfan
  /.Pfa /pfan
  /.PfA /pfan
  /.PFa /pfan
  /.PFA /pfan
>> readonly def

% Check whether the file name has pfa or pfb extension.
/pfa_or_pfb? {      % s -> false   |    /name true
  dup length 4 lt {
    pop //false
  } {
    dup length 4 sub 4 getinterval //pfa_pfb_dict exch .knownget
  } ifelse
} bind def

% pf[ba]filename makeafm -
/makeafm {
  count 0 eq {(Missing font file name\n) printquit} if
  /ifn edef
  ifn length 0 eq {(Empty font file name\n) printquit} if
% the following piece of the code does, in fact, the job of a system shell,
% i.e., it analyses the supplied names, appends extensions if needed,
% and check files:
  /pfbn () def /pfan () def /pfmn () def % initialisation
  [ t1_glyph_equivalence { pop } forall ] { % disable glyph substitution
    t1_glyph_equivalence exch undef
  } forall
  ifn pfa_or_pfb? {
      ifn dup length string copy def
      ifn dup length 4 sub 0 exch getinterval /ifn edef
  } if
  pfbn () eq pfan () eq and dup {% no extension was supplied, try ".pfb"
    /pfbn ifn (.pfb) concatstrings def
  } if
  pfbn () ne {% check whether "filename.pfb" exists
    pfbn status {pop pop pop pop /isPFB true def}{/pfbn () def} ifelse
  } if
  pfbn () eq and {% checking "filename.pfb" unsuccessful, try ".pfa"
    /pfan ifn (.pfa) concatstrings def
  } if
  pfan () ne {% check whether "filename.pfa" exists
    pfan status {pop pop pop pop /isPFB false def}{/pfan () def} ifelse
  } if

  pfbn () eq pfan () eq and {
    (Neither pfa nor pfb found\n) printquit
  } if

  /ofn ifn (.afm) concatstrings def
  ofn status {
    pop pop pop pop (Resulting file exists\n) printquit
  } if
  /ofi ofn (w) file def

  /pfmn ifn (.pfm) concatstrings def
  pfmn status {
    pop pop pop pop
  }{
    () pfmn {
      (/) search dup not { pop (\\) search } if {
        4 -1 roll exch concatstrings exch concatstrings exch
      }{
        exit
      } ifelse
    } loop
    (pfm/) exch concatstrings concatstrings
    dup status {
      pop pop pop pop /pfmn edef
    }{
      pop /pfmn () def (pfm file not found -- ignored\n) print
    } ifelse
  } ifelse
  //systemdict /.setsafe known {
    <<
      /PermitFileReading
        [ pfmn dup length 0 eq { pop } if
          isPFB {pfbn}{pfan} ifelse
        ]
      /PermitFileWriting [ ]
      /PermitFileControl [ ]
    >> setuserparams
    .locksafe
  } if

  pfmn
  isPFB {pfbn}{pfan} ifelse
  printafm

} def

% Check for command line arguments.
[ shellarguments
  { ] dup length 1 eq {
       0 get makeafm
     }{
       (This is PF2AFM -- AFM generator \(ver. 1.00\)\n) print
       (Usage: gs [-dNODISPLAY] -- pf2afm.ps disk_font_name\n) printquit
     } ifelse
  }
  {pop}
ifelse